Original post by Peter Mckean - MD, Synthetix
You can't have missed the announcements from Facebook and Microsoft that they have developed 'bot' frameworks that will enable a whole new category of self-service applications running within platforms such as Facebook Messenger or Skype. These were followed by near-hysterical speculative articles proclaiming the death of the call-centre.
Not quite true.
I have a personal stake in this game, and a perspective on why bots aren't the be-all and end-all of customer service - but also where they can be useful.
In 2001, Synthetix created one of the first commercially available customer-service Virtual Agent bot platforms - SmartAgent. It was clever - able to remember context, collect information and post it through databases, chat with customers and, crucially, answer their questions. We found customers for the software with the BBC, Ford and BT amongst many. It was an exciting time and we believed we were at the vanguard of an AI revolution, convinced by reports from excitable industry analysts predicting an explosion in bots.
It never happened.
One problem was that the analysts talked to vendors such as us, all of whom painted a rosy picture of the market potential. Then, they wrote their reports and we read them, literally believing our own hype. Another problem was that the product category was difficult to do well. Without exception, Virtual Agents required a big investment in time to get right, from the visuals to the conversational elements to the customer journey. And choosing a visualisation (a friendly human face) that worked for all markets was difficult, if not impossible.
No, Virtual Agents were a perfect niche product - always the bridesmaid, never the bride. Always threatening to become mainstream, never succeeding. Finally realising that a focus on Virtual Agents would always inhibit our growth potential, we radically changed strategy, taking our core expertise in knowledge-base management and building a portfolio of interconnected customer service tools such as live chat, email management and FAQ web self-service which put a knowledge-base at the heart of a multi-channel customer experience. The hardest decision was not to consider the bots as the centrepiece of the portfolio or the ultimate solution to any client problem, but as just another service no more or less valid that the other tools we developed. It was the right decision. We've grown massively on the back of that change in business model and Virtual Agents now represent just 5% of our business.
Of course, technologies like Siri and Cortana have renewed interest in Virtual Agents. Strictly speaking, Siri and company are Virtual Assistants, not Virtual Agents - digital PAs, not digital customer-service agents. The new bot frameworks from Facebook and Microsoft blur the lines, somewhat, allowing users to self-serve through a conveniently available interface. Beyond the natural language extensions, these are simply channels - new ways for customers to contact brands and receive support directly inside environments such as Facebook Messenger. If you have an online presence and a healthy customer-base, it's your responsibility to support your customers however they choose to get in touch - there's no point having great telephone support if most of your customers visit your website, for instance. In fact, statistics we collect from our live chat instances suggest that 13% of customers simply would not contact you, except through online channels!
Companies like Synthetix will connect these new channels to our existing automated natural language self-service engines and offer escalation up to live agents when it will support a good customer experience. Ultimately, these new developments are not a threat to call centres, they're a welcome opportunity to serve customers better and relieve agents of having to deal with humdrum queries which are more appropriately handled through self-service mechanisms.